Home rehab's accuracy gap is calibration, not cameras
Remote motion-tracking already agrees with the clinic on relative movement. It is off by 18 degrees on the absolute number — and that single bias is what keeps it out of clinical use.
The call, up front. The sensor was never the problem. Home-based rehab motion tracking already tracks change as well as a clinician — but a systematic 18-degree bias in the absolute angle is what blocks clinical-grade adoption. The gap is a calibration layer, not a better camera.
The gap
A Kinect-based system that agrees with the gold standard at ICC 0.89 is already good enough to track a patient’s progress. What it cannot do is state a defensible absolute angle — and reimbursement, software-as-a-medical-device classification, and clinician trust all hang on the absolute number.
Source: GAPTIQ engine — challenge definition; validation vs manual goniometry (ICC = 0.89, 18° bias)
Spend on a per-home calibration protocol, not a better sensor. The relative-accuracy problem is already solved.
It can already see the movement. It just can’t yet name the angle.
GAPTIQ Signal · Jun 2026
So what
The whitespace is a calibration-and-validation layer that turns a consumer-grade signal into a SaMD-defensible measurement — sold to the digital-rehab platforms that already have the cameras and the patients. Whoever owns the calibration owns the clinical claim.
Source: Usability and Reliability of a New Physical Therapy Tool, ClinicalTrials.gov. Surfaced by the GAPTIQ engine.